Work

AI-Driven Detection of Unauthorized Buildings to Protect Hamlet Heritage in Fragile Territories in Italy

Year: 2025

Type: article

Abstract: . Unauthorised construction continues to endanger the integrity of historic hamlets and vernacular heritage in fragile Italian territories. This research presents an automated methodology for detecting unregistered buildings through the integration of cadastral datasets, high-resolution orthophotos, and the Segment Anything Model (SAM), a foundation model for image segmentation. By prompting SAM with cadastral centroids, building footprints were extracted and compared to official records to identify spatial discrepancies and undocumented structures. The method proved effective in detecting both geometric anomalies and potential building code violations. Despite some limitations related to material reflectivity and complex roof morphologies, the workflow is reproducible, scalable, and open source. By combining AI-assisted segmentation and GIS-based spatial analysis, the approach contributes to the development of digital tools for heritage documentation, territorial monitoring, and planning control.

Source: The international archives of the photogrammetry, remote sensing and spatial information sciences/International archives of the photogrammetry, remote sensing and spatial information sciences

DOI: 10.5194/isprs-archives-xlviii-m-9-2025-1349-2025

Authors: Marco Seccaroni, Domenico D’Uva

Institution: Politecnico di Milano

Language: English

Topic: Automated Road and Building Extraction

Subfield: Ocean Engineering

Field: Engineering

Domain: Physical Sciences

SDGs: Sustainable cities and communities, Industry, innovation and infrastructure

Open Access status: diamond

OpenAlex ID: W4414798078